Abstracts

RATIONALE: Emergent EEG (E-EEG) should be ordered if it is expected to have a major impact on patient diagnosis and management. Few published data suggest misuse of the test. Our objective was to identify the reasoning why the test was ordered and through correlation with the test results determine its suitability for the clinical situation encountered.
METHODS: We retrospectively reviewed all E-EEGs (ordered to be performed within 1 hour) during the last 2 years in our University Hospital. We collected the following data: demographics, reason of admission, service ordering the test, question to be answered, type of E-EEG (routine or prolonged), presence of suspicious clinical signs (SCS) including involuntary movements or blank starring, previous history of seizures, antiepileptic medications and interpretation. Univariate and multivariate analysis was performed using the chi-square test and logistic regression ([alpha]= 0.05).
RESULTS: One hundred twenty six E-EEGs were performed in 119 patients (62 M and 57 F, mean age 56 years). Clinical history revealed mental status change (MS) in 28 (22.2%), cardiac or respiratory arrest or prolonged hypotension (CRAH) in 18 (14.3%), systemic primary disorder (SD) in18 (14.3%), epilepsy in 16 (12.7%), stroke in 15 (11.9%) and tumor in 13 (10.3%) cases. Sixty six (55%) E-EEGs were ordered by the neurology or neurosurgery and 32 (25.4%) by the Medical Intensive Care Unit (MICU) service. Status epilepticus (SE) had to be ruled out in 61% of cases, seizures in 18%, encephalopathy in 10% and non-convulsive status epilepticus (NCSE) in 8.7%. SCS before the E-EEG was observed in 62% of cases with question of SE and in 61% with seizure exclusion. SCS was observed in 36% of cases with NCSE exclusion. E-EEG was done to exclude SE in 83% of CRAH cases, in 69% of cases with epilepsy and 63% with MS changes. In only 11% of MS changes, E-EEG was done to exclude NCSE. Prolonged E-EEG was requested in only 4 cases (3.2%). Encephalopathy was confirmed by E-EEG when the test was done to rule it out (odds ratio 5.4, 95% CI 1.1-26, P [lt] 0.05). No such association was found between reasoning for ordering the test and confirmation of the suspicion by E-EEG results for SE or NCSE. With recent SCS, SE or NCSE was found more often on E-EEG (5.6, 1.2-26, P [lt] 0.05), but not epileptiform activity or electrographic seizure (EAES) or encephalopathy. In a logistic regression analysis only CRAH was associated with SE or NCSE (5.3, 1.4-20, P [lt] 0.05), tumor with EAES (5.9, 1.6-21.3, P [lt] 0.01) and tumor or history of seizures with encephalopathy (0.08, 0.01-0.7 and 0.3, 0.1- 0.9, P [lt] 0.05 respectively).
CONCLUSIONS: Our findings suggest that E-EEG is ordered more frequently in patients with MS changes, CARH or SD, by neurologists or neurosurgeons, to rule out SE. SCS were observed in only two thirds of cases when the test was ordered to exclude SE and in one third of cases to exclude NCSE, questioning the suspicion level of the ordering physician. Prolonged EEG, a more appropriate test according to the clinical situation and question asked, was ordered infrequently. E-EEG will most likely reveal SE or NCSE, if done post CRAH, and EAES in the presence of tumor. Encephalopathy is less likely to be found with a history of seizures or tumor.